We owe our sense of smell to the ability of odorant compounds to activate odorant receptors present on the membrane surface of olfactory sensory neurons in our nose. Our present knowledge indicates that, in mammals, there are around 1,000 types of odorant receptors and that any given neuron only expresses one type of receptor, to which different odorants can bind with different affinities. Thus, any odorant can be thought of as being "encoded" by the olfactory system in the form of a set of affinities between itself and each of the odorant receptors. Conversely, odorant receptors may, in principle, be categorized by their pattern of activation by (some set of) odorants. Present approaches to detect matches between odorants and their receptors range from in-vivo patch-clamp recordings from the olfactory epithelium, bulb or cortex, to Ca2+ imaging of a small number of cultured neurons, in response to a small set of odorants. In addition, since the neuron's response is concentration-dependent, each odorant must be screened at various concentrations with isolated olfactory sensory neurons. Once the response of a given neuron by a given odorant at any concentration is detected, the receptor can be cloned by reverse-transcriptase polymerase chain reaction and subsequently sequenced. Although this approach allows for exquisite detail in characterizing the molecular structure of the receptor, it yields an incomplete functional description, i.e. it might bind to hundreds of other odorants that had to be missed in the experiment due to throughput limitations. Furthermore, the odorant receptor can only be probed once because the responsive cell is probed at random. Additionally, the approach is inadequate for performing an exhaustive search of all the odorant receptors that are activated by a given odorant from a library of hundreds of odorants; hence the encoding of the odorant within the olfactory system is unknown. We seek to obtain a functional description of all the odorant receptors present in the mouse olfactory epithelium. This functional description will not require knowledge of the receptors' molecular structure. We propose to develop a high-throughput experimental system that will allow for detecting matches between a large library of around 100 odorants and the complete set of interacting receptors (approximately 1,000), i.e. we will find the exact encoding of each member of the odorant library. The proposed system consists of 1) a computer-controlled microfluidic device that can perfuse a large microarray of olfactory sensory neurons with sequences of odorants; 2) an optics setup capable of performing Ca2+ imaging over large areas (>5,000 cells); and 3) simple image/data processing algorithms for automated extraction of odorant-receptor matches.